摘要
机械臂运动轨迹容易受到外界不确定因素的干扰,导致运动轨迹跟踪误差较大,振动现象较为严重,不能很好地满足机械臂的准确定位.建立了双关节机械臂模型简图,采用RBF(径向基函数)神经网络算法非线性积分滑模控制机械臂的运动轨迹.分析了RBF神经网络算法结构,推导了RBF神经网络算法非线性积分滑模控制方程式和在线补偿法则,引用李雅普诺夫函数证明机械臂控制系统的稳定性.采用Matlab软件对双关节机械臂运动轨迹跟踪误差进行仿真,并与PID控制系统的跟踪误差进行对比和分析.仿真误差曲线显示,机械臂运动轨迹在受到外界干扰因素的影响时,采用RBF神经网络算法非线性积分滑模控制方法,不仅跟踪误差较小,而且输入转矩波动幅度较小.机械臂末端采用RBF神经网络算法非线性积分滑模控制方法,提高了机械臂的定位精度,降低了抖动幅度.
The trajectory of the mechanical arm is easy to be disturbed by the uncertainties of the outside world,which leads to the large error of motion trajectory tracking,and the vibration phenomenon is more serious,which can not satisfy the accurate positioning of the mechanical arm.A schematic diagram of the bijoint mechanical arm model is established,and the motion trajectory of the manipulator is controlled by the nonlinear integral sliding mode of radial basis function（RBF）neural network.Analyzing the structure of RBF neural network algorithm,and RBF neural network nonlinear integral sliding mode control algorithm was deduced equation and online compensation principle,reference Lyapunov function to prove the stability of the mechanical arm control system.Matlab software is used to simulate the tracking error of the motion trajectory of the double-joint manipulator,and the tracking error of the PID control system is compared and analyzed.Simulation error curve show that the mechanical arm trajectory when under the influence of interference factors,by using RBF neural network algorithm for nonlinear integral sliding mode control method,not only the tracking error is small,and the input torque fluctuation amplitude is smaller.The nonlinear integral sliding mode control method of RBF neural network is used to improve the positioning accuracy of the manipulator and reduce the vibration amplitude.
作者
董君
陈立
DONG Jun;CHEN Li(School of Public Education,Jilin Province Economic Management Cadre Institute,Changchun 130012,China;School of Mechanical Engineering and Automation, Fuzhou University, Fuzhou 350108, China)
出处
《中国工程机械学报》
北大核心
2018年第2期106-110,共5页
Chinese Journal of Construction Machinery
基金
福建省自然科学基金资助项目(2013J01011)
关键词
机械臂
RBF神经网络算法
非线性积分
滑模控制
跟踪误差
仿真
manipulator
RBF neural network algorithm
nonlinear integral
sliding mode control
tracking error
simulation
